Poster Presentation/Demo Abstract
Need to find a specific segment in an hour-long web video of a lecture? Want to use small segments of video for teaching and learning? The SpokenMedia project will demonstrate our automatic lecture transcription solutions, and will discuss a number of tools/services to aid educators/learners in using this more granular media.
Statement of the problem or issue
The production and use of video lectures for teaching and learning continues to accelerate around the world. As the collections of teaching and learning video grow, learners and educators are challenged to find the video resources they need, at the granularity levels that are useful to them. Search and retrieval of video and other rich media are limited by the metadata that is used to describe the video, yet what if the topic the user needs to find isn’t described in the metadata?
Importance or relevance to other faculty, staff, students, departments, and programs
SpokenMedia lecture transcription can provide a searchable transcript and an 80-85% accurate caption for all lecture-style video at MIT, all done automatically. Not only does this dramatically increase the potential to find relevant whole videos, but with project developed tools users will be able to locate and use specific relevant segments for their teaching, learning or research.
Brandon Muramatsu, MIT
Andrew McKinney, MIT
(Presented at MIT Educational Technology Fair 2009)
Topic Area(s)
2. Finding and integrating digital content into the curriculum